Readmission prevention efforts have become "a priority in the era of health care reform" (Stevens, 2015, p. 123). Unplanned 30-day hospital readmissions, according to Fluitman et al. (2016), are prevalent and costly. McIlvennan et al. (2015) report that readmissions are associated with unfavorable patient outcomes and that efforts to address this issue have become a chief focus among hospitals and other health care organizations such as skilled nursing facilities (SNFs), rehabilitation facilities, and home health care agencies.
Despite research findings demonstrating the impact of hospital readmissions on patient outcomes and their costly burden to the health care system, data from the Center for Health Information and Analysis (2020a, 2020b) report that unplanned, all-payer hospital readmission rates increased from 15.2% in 2013 to 16.1% in 2017. Over the same period, the unplanned, all-payer readmission rates for the study hospital increased from 9.9% to 10.4% (Winchester Hospital All-payer unplanned readmission rates, 2013, 2017).
The purpose of this study is to explore unplanned 30-day hospital readmissions at a 229-bed community hospital 8 miles north of Boston. Using the RCA process, the aim will be twofold:
* First, identifying the root cause of unplanned readmissions for patients who return to the hospital within 30 days of discharge.
* Second, the use of the data gathered from this study that may potentially change and/or refine nursing practices and interventions to improve patient outcomes, reduce the number of readmissions, and decrease associated health care costs.
Background Information/Significance/Scientific Rationale
Certain comorbidities are associated with an increased risk for readmission. According to Silverstein et al. (2008), these comorbidities include cancer, renal failure, paralysis, diabetes with chronic complications, heart failure, liver disease, and chronic obstructive pulmonary disease. Koekkoek et al. (2011) find common reasons for readmission to be dementia/delirium/altered mental status, pneumonia, urinary tract infection, and using greater than 10 and/or greater than three new medications at time of discharge. Silverstein et al. (2008) report that other characteristics predictive of readmission include being 75 years or older, male, and African American; being treated in a medical versus surgical service, having Medicare-only insurance, and being discharged to an SNF. According to Stevens (2015), hip and knee arthroplasty were added to the 30-day readmission risk list in 2015 by the Centers For Medicare & Medicaid Services (2019) in their Hospital Readmission Reduction Program.
In addition to these identified above readmission characteristics, Boutwell et al. (2017) have coined the term "super utilizers" meaning patients who are frequent users of the acute care setting. This term, super utilizers, adds additional burden to the health care system. Boutwell et al. suggest that rather than viewing these patients as "complex," "difficult," or "nonadherent," high utilization should be viewed as "a symptom of unmet needs" (p. 18) and utilization for these patients is driven by one or a combination of needs: medical, behavioral, and social (Boutwell et al., 2017, p. 30). It is important to consider all three categories when performing RCA so that a primary diagnosis, chief complaint, or complicated medical history is not mistaken as the root cause of an unplanned readmission. "It is the human, individual reason that this person, with all his/her complexities and social needs comes to the hospital so frequently, while another person who is similar does not frequent the hospital" (Boutwell et al., 2017, p. 19).
RCA (Root Cause Analysis)
Peerally et al. (2017) call the RCA process "a method of structured risk identification and management in the aftermath of adverse events" (Peerally et al., 2017, p. 417). RCA, according to Lean Six Sigma Definition (n.d.) and Serrat (2010), was invented by Sakichi Toyoda for use in his family's manufacturing company, which his son would eventually develop into Toyota Industries Corporation. Sakichi was convinced that every problem stems from a single root cause, which can be determined by asking the question "Why?" five times. Once the root cause is identified, a solution can then be applied. The goal of an RCA is to prevent the recurrence of a problem (Ouslander et al., 2016; Uberoi et al., 2007) using this solution. As Uberoi et al. suggest, the RCA approach may ultimately help to reduce the number of readmissions and the associated costs.
Percarpio et al. (2008) say that RCA is an analysis framework used in health care to determine systemic causes and prevent recurrences of adverse events. Iedema et al. (2006) reported that the Veterans Administration (VA) trialed RCA at four of its hospitals and went system-wide with it in 2000. The VA then took it a step further by using RCA, not only for clinical errors, but for "close calls" as well. In 2001, as reported by Neal et al. (2004), the Department of Health updated its publication: Building a Safer NHS for Patients to include "a plan to build expertise within the National Health Service in the technique of root cause analysis" (p. 75). The Joint Commission allows organizations to use assessment tools that they consider appropriate to achieve an outcome that will mitigate or eliminate risk with RCA given as the first example of tools that might be used (The Joint Commission, 2020)
Participant Selection/Eligibility
The Readmission Prevention Team consists of three registered nurses who identify patients at high risk for readmission based on the noted below inclusion and exclusion criteria. A relationship is established by meeting the patient in the hospital and participating in patient care rounds. The readmission prevention nurse collaborates with the patient and caregivers, primary nurse, case manager, and hospitalist to facilitate an appropriate discharge plan. Any participant who has experienced a nonplanned readmission to the study hospital within 30 days of discharge is eligible for enrollment in the study.
Inclusion Criteria
* Admitted to the study hospital
* Enrolled in the Readmission Prevention Program
* Discharged from the study hospital to home, home with home care, SNF, or acute rehabilitation hospital
* Readmitted to the study hospital within 30 days of discharge
Exclusion Criteria
* Under the care of an Accountable Care Organization provider and/or Atrius provider
* Primary diagnosis of a behavioral health condition
* Admitted to the obstetrical service
* Returning to the hospital for a planned procedure or surgery
* Under observation status
Subject Enrollment/Consent Process
All patients who meet the inclusion criteria and have had a "5 Why RCA" worksheet completed on readmission from March 2018 to March 2020 will be included in the study. This worksheet will be coded with a study number to facilitate data analysis. To ensure anonymity, no patient identifiers were included on the worksheet. Participant risk, discomfort, and consent do not apply because this is a retrospective worksheet audit.
Study Design/Procedures
Using a descriptive, correlational, qualitative design, demographic data will be collected on each subject including age, gender, days between discharge and readmission, and the venue from which the patient was admitted (Appendix A). The "5 Why RCA" (Appendix B with example), according to Bialek et al. (n.d.) is a questioning process designed to investigate the details of a problem. It includes a problem statement and five numbered lines in descending order for the responses to the "Why." For data analysis, the last Why, as determined by the RCA, will be categorized as either medical, behavioral, and/or social.
Study Calendar/Schematic/Schedule
Following institutional review board approval, the 5 Whys worksheets completed at the study hospital from March 2018 to March 2020 were reviewed by the Readmission Prevention Team. Categorization of the last Why was done independently by each member of the Readmission Prevention Team applying the "category scheme procedure." This procedure, defined by Polit and Beck (2006), involves "careful reading of the data with an eye to identifying underlying concepts and clusters of concepts." Categories are developed based on scrutiny of the actual data. The category scheme "may focus on actions or events or on different phases in a chronologic unfolding of an experience. In developing it, related concepts are often grouped together to facilitate the coding process" (p. 384). For the purpose of this study, categories will be defined as medical, behavioral, and social. This particular scheme was described by Boutwell et al. (2017) in the New York State Department of Health Series Program Final Report: Improving care for Super Utilizers (2017).
Potential Risks and Discomforts
There are no risks or discomforts for subjects with this study, as only historic data will be collected from the "5 Whys" worksheet.
Potential Benefits
Potential benefits are changes in practice/protocols/interventions that could arise from the findings of the study. Such changes in practice could lead to better outcomes for patients at high risk for readmission and lower 30-day hospital readmission rates.
Significance to Nursing/Case Managers
Nurses/case managers are on the front line in all patient care domains. Readmission prevention nurses and case managers are charged with decreasing the number of unplanned hospital readmissions. The results of this study may help identify variables that contribute to these problems and show us the way to making them a part of the past.
Statistical Analysis
The coded data from the last Why were described by frequency and grouped by the following demographic criteria: age, gender, days between discharge and readmission, and the discharge venue. The frequencies of each group were statistically compared through one Pearson's chi-squared test for each demographic. Significance in the chi-squared test indicates the frequency of a coded readmission reason differs from an expected frequency based off the number of individuals in that group. If significance is found, an effect size will be calculated to determine the size and direction of the difference.
Data Management
All data were anonymous, maintained in a locked file cabinet, and accessible only by the investigators.
Data and Safety Monitoring and Quality Assurance
The primary investigator (PI) will assume responsibility for ensuring that data are maintained in a locked file cabinet located in the office of the investigator and accessible only to the PI and coinvestigators.
Results
A total of 150 RCA worksheets met the study criteria and were reviewed by the Readmission Prevention Team both individually and then collectively.
Of the patients readmitted,
* 70% were females
* 30% were males
The age decades of those readmitted were the 70s and 80s as shown in Figure 1.
The discharge venues from where the patients were readmitted are noted in Figure 2.
The primary root cause categories of readmissions are demonstrated in Figure 3:
* medical causes (92%)
* behavioral health causes (17.3%)
* social root causes (6.7%).
A total of 19 (13%) of the readmissions showed multiple root causes:
* 11 had both medical and behavioral health root causes
* 2 had both medical and social root causes
* 1 had both behavioral health and social root causes
* 5 had all three root causes-medical, behavioral, and social
The days between the discharge and the readmission (see Figure 4) show a slight bell curve shape, which can be seen centering on approximately 2 weeks (mean = 13.89 days).
Examining each data set with a chi-square contingency analysis test, nonrandom trends were found in gender ([chi]2 = 24, df = 1, n = 150, p <= .001), age ([chi]2 = 169.95, df = 6, n = 150, p <= .001), and venue ([chi]2 = 237.33, df = 6, n = 150). The days between discharge and readmission showed a strong yet nonsignificant nonrandom trend ([chi]2 = 24.507, df = 15, n = 150, p = .057).
When the data on each readmission category (medical, behavioral, and social) are reviewed, some of the trends seen in the aggregate data are repeated with minor differences. In each readmission cause, the same gender and age trends are seen; however, behavioral readmissions show an overwhelming majority of patients in their 70s and few in their 80s, differing from the other two readmission causes. Another difference is in the social readmission venue area where patients were readmitted equally from home without services and home with home care and in the days between readmissions. These differences may be the result of the small number of social readmissions in the data set.
Discussion Section
The major root cause of unplanned readmissions for patients who returned to the hospital within 30 days was medical. This finding is consistent with that found in the literature-a medical exacerbation is the prominent readmission cause.
Although behavioral health and social root causes were identified, they were not as individually predictive of readmissions as the medical root cause. Nineteen of the subjects (13%) had multiple root causes for readmission. Of these 19 subjects, 11 had both medical and behavioral health root causes whereas five subjects had all three root causes. These findings indicate that the medical treatment plan for the patient on discharge needs to continue to be a major focus, but other aspects of the discharge plan should address the behavioral health and social needs of the patient.
The common venue from which the subject returned to the hospital was "home with home care." This seems to indicate the patient needed more intensive care in the community than the home care clinicians could provide. Each patient is unique in their health care needs, and the discharge plan developed in the hospital takes these needs into consideration. Determinants of the patient's care can be medical, functional, and social (Alper et al., 2020) with the health care team assessing how best to meet these needs on discharge.
Occasionally, a patient may be borderline for home care but is insistent on going home. In these cases, developing a backup discharge plan or Option B may be needed. Prior to the patient being discharged home with home care, the patient can also be screened for an SNF. By having this backup venue for the patient, rehospitalization might be avoided. An important component of Option B is that it is clearly communicated to the patient/family, home care agency, and SNF.
The second most common venue from which subjects were readmitted was the SNF. This suggests that there might be an opportunity for these readmissions to be reviewed by both the hospital and the SNF to find whether there was a way to prevent them from happening. Communication that could avoid these readmissions might include the SNF health care providers calling the hospital emergency department (ED) to discuss the reason the patient is coming and what plan might help to return the patient to the SNF instead of admitting to the hospital. Important information to know by all involved would be the skills/diagnostic interventions that the SNF has in place to deal with the patient's medical needs. Another innovative practice might be to develop a meeting of SNF administrators and appropriate hospital personnel to review these admissions to develop strategies to prevent them in the future.
The days between discharge and readmission show a bell curve from day 1 through days 19-20. This data confirms the need to monitor a patient's health intensely through the 30-day window and does show the need to have medical follow-up appointments in a timely fashion.
Implications for Case Managers
The results of this study demonstrated that the majority of readmissions were related to the medical diagnosis, that the majority of patients were in their 70s and 80s, and that the readmission time frame was within 2 weeks of the initial discharge. Case managers and readmission prevention nurses already had strong processes in place for these discharge plans and follow-up.
Study results also show where more emphasis and creativity can be employed to potentially reduce further readmissions. Although the medical category was the predominant readmission cause, behavioral and social determinants of health also played a part in readmissions. These behavioral and social causes of readmissions should be reviewed to assess whether the hospital can develop or strengthen processes that might lessen their impact on readmissions.
The most frequent readmission venue was home with home care, which seems to indicate that the patient needed more care than what could be provided at home. One creative way of preventing this might be the development of Option B. If there was some question that home care might not sustain the patient at home, the patient could be screened for an SNF prior to discharge so that there would be a backup venue for the patient other than going to the hospital ED.
The second most common readmission venue was the SNF. More communication between the SNF and the hospital ED might reduce the number of readmissions. A process where the SNF nurse practitioner/physician calls the hospital ED to discuss the patient's condition could expedite the patient's care and reduce the number of readmissions to the hospital. Along with this process, developing an SNF skill capability list could help the hospital's ED providers understand the care that a patient can receive in the SNF. In addition, direct, regular communication could be established between SNF administrators and appropriate hospital personnel for more open dialogue on creating ways that might prove helpful to reduce readmissions.
References